Built-In Zoo Models#
This page lists all of the natively available models in the FiftyOne Model Zoo.
Check out the API reference for complete instructions for using the Model Zoo.
AlexNet model architecture from "One weird trick for parallelizing convolutional neural networks" trained on ImageNet
CenterNet model from "Objects as Points" with the Hourglass-104 backbone trained on COCO resized to 1024x1024
CenterNet model from "Objects as Points" with the Hourglass-104 backbone trained on COCO resized to 512x512
CenterNet model from "Objects as Points" with the MobileNetV2 backbone trained on COCO resized to 512x512
CenterNet model from "Objects as Points" with the ResNet-101v1 backbone + FPN trained on COCO resized to 512x512
CenterNet model from "Objects as Points" with the ResNet-50-v1 backbone + FPN trained on COCO resized to 512x512
CenterNet model from "Objects as Points" with the ResNet-50v2 backbone trained on COCO resized to 512x512
Hugging Face Transformers model for image classification
CLIP text/image encoder from "Learning Transferable Visual Models From Natural Language Supervision" trained on 400M text-image pairs
DeepLabv3+ semantic segmentation model from "Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation" with Xception backbone trained on the Cityscapes dataset
DeepLabv3+ semantic segmentation model from "Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation" with MobileNetV2 backbone trained on the Cityscapes dataset
DeepLabV3 model from "Rethinking Atrous Convolution for Semantic Image Segmentation" with ResNet-101 backbone trained on COCO
DeepLabV3 model from "Rethinking Atrous Convolution for Semantic Image Segmentation" with ResNet-50 backbone trained on COCO
Densenet-121 model from "Densely Connected Convolutional Networks" trained on ImageNet
Densenet-161 model from "Densely Connected Convolutional Networks" trained on ImageNet
Densenet-169 model from "Densely Connected Convolutional Networks" trained on ImageNet
Densenet-201 model from "Densely Connected Convolutional Networks" trained on ImageNet
Hugging Face Transformers model for monocular depth estimation
Hugging Face Transformers model for object detection
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-B/14 distilled
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-B/14 distilled
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-g/14
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-g/14
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-L/14 distilled
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-L/14 distilled
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-S/14 distilled
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-S/14 distilled
EfficientDet-D0 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 512x512
EfficientDet-D0 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO
EfficientDet-D1 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 640x640
EfficientDet-D1 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO
EfficientDet-D2 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 768x768
EfficientDet-D2 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO
EfficientDet-D3 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 896x896
EfficientDet-D3 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO
EfficientDet-D4 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 1024x1024
EfficientDet-D4 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO
EfficientDet-D5 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 1280x1280
EfficientDet-D5 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO
EfficientDet-D6 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 1280x1280
EfficientDet-D6 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO
EfficientDet-D7 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 1536x1536
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" atrous version with Inception backbone trained on COCO
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" atrous version with low-proposals and Inception backbone trained on COCO
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with Inception v2 backbone trained on COCO
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with NAS-net backbone trained on COCO
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with low-proposals and NAS-net backbone trained on COCO
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with ResNet-101 backbone trained on COCO
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with low-proposals and ResNet-101 backbone trained on COCO
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with ResNet-50 backbone trained on COCO
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with ResNet-50 FPN backbone trained on COCO
Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with low-proposals and ResNet-50 backbone trained on COCO
FCN model from "Fully Convolutional Networks for Semantic Segmentation" with ResNet-101 backbone trained on COCO
FCN model from "Fully Convolutional Networks for Semantic Segmentation" with ResNet-50 backbone trained on COCO
GoogLeNet (Inception v1) model from "Going Deeper with Convolutions" trained on ImageNet
Hugging Face Transformers model for zero-shot semantic segmentation
Inception v2 model from "Rethinking the Inception Architecture for Computer Vision" trained on ImageNet
Inception v3 model from "Rethinking the Inception Architecture for Computer Vision" trained on ImageNet
Inception v4 model from "Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning" trained on ImageNet
Keypoint R-CNN model from "Mask R-CNN" with ResNet-50 FPN backbone trained on COCO
Mask R-CNN model from "Mask R-CNN" atrous version with Inception backbone trained on COCO
Mask R-CNN model from "Mask R-CNN" with Inception backbone trained on COCO
Mask R-CNN model from "Mask R-CNN" atrous version with ResNet-101 backbone trained on COCO
Mask R-CNN model from "Mask R-CNN" atrous version with ResNet-50 backbone trained on COCO
Mask R-CNN model from "Mask R-CNN" with ResNet-50 FPN backbone trained on COCO
Fine-tuned SAM2-hiera-tiny model from "Medical SAM 2 - Segment Medical Images as Video via Segment Anything Model 2"
MNASNet model from "MnasNet: Platform-Aware Neural Architecture Search for Mobile" with depth multiplier of 0.5 trained on ImageNet
MNASNet model from "MnasNet: Platform-Aware Neural Architecture Search for Mobile" with depth multiplier of 1.0 trained on ImageNet
MobileNetV2 model from "MobileNetV2: Inverted Residuals and Linear Bottlenecks" trained on ImageNet
MobileNetV2 model from "MobileNetV2: Inverted Residuals and Linear Bottlenecks" trained on ImageNet
Hugging Face Transformers OmDet-Turbo
OPEN CLIP text/image encoder from "Learning Transferable Visual Models From Natural Language Supervision" trained on 400M text-image pairs
Hugging Face Transformers OWL-ViT
ResNet-50 v1 model from "Deep Residual Learning for Image Recognition" trained on ImageNet
ResNet-50 v2 model from "Deep Residual Learning for Image Recognition" trained on ImageNet
ResNet-101 model from "Deep Residual Learning for Image Recognition" trained on ImageNet
ResNet-152 model from "Deep Residual Learning for Image Recognition" trained on ImageNet
ResNet-18 model from "Deep Residual Learning for Image Recognition" trained on ImageNet
ResNet-34 model from "Deep Residual Learning for Image Recognition" trained on ImageNet
ResNet-50 model from "Deep Residual Learning for Image Recognition" trained on ImageNet
ResNeXt-101 32x8d model from "Aggregated Residual Transformations for Deep Neural Networks" trained on ImageNet
ResNeXt-50 32x4d model from "Aggregated Residual Transformations for Deep Neural Networks" trained on ImageNet
RetinaNet model from "Focal Loss for Dense Object Detection" with ResNet-50 FPN backbone trained on COCO
R-FCN object detection model from "R-FCN: Object Detection via Region-based Fully Convolutional Networks" with ResNet-101 backbone trained on COCO
RT-DETR-l model trained on COCO
RT-DETR-x model trained on COCO
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model 2 (SAM2) from "SAM2: Segment Anything in Images and Videos"
Segment Anything Model (SAM) from "Segment Anything" with ViT-B/16 backbone trained on SA-1B
Segment Anything Model (SAM) from "Segment Anything" with ViT-H/16 backbone trained on SA-1B
Segment Anything Model (SAM) from "Segment Anything" with ViT-L/16 backbone trained on SA-1B
Hugging Face Transformers model for semantic segmentation
ShuffleNetV2 model from "ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design" with 0.5x output channels trained on ImageNet
ShuffleNetV2 model from "ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design" with 1.0x output channels trained on ImageNet
Hugging Face Transformers model for zero-shot image classification
SqueezeNet 1.1 model from "the official SqueezeNet repo" trained on ImageNet
SqueezeNet model from "SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and" trained on ImageNet
Inception Single Shot Detector model from "SSD: Single Shot MultiBox Detector" trained on COCO
Single Shot Detector model from "SSD: Single Shot MultiBox Detector" with MobileNetV1 backbone trained on COCO
MobileNetV1 model from "MobileNetV2: Inverted Residuals and Linear Bottlenecks" resized to 640x640
FPN Single Shot Detector model from "SSD: Single Shot MultiBox Detector" with MobileNetV1 backbone trained on COCO
MobileNetV2 model from "MobileNetV2: Inverted Residuals and Linear Bottlenecks" resized to 320x320
FPN Single Shot Detector model from "SSD: Single Shot MultiBox Detector" with ResNet-50 backbone trained on COCO
VGG-11 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" with batch normalization trained on ImageNet
VGG-11 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" trained on ImageNet
VGG-13 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" with batch normalization trained on ImageNet
VGG-13 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" trained on ImageNet
VGG-16 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" with batch normalization trained on ImageNet
VGG-16 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" trained on ImageNet
VGG-16 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" trained on ImageNet
VGG-19 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" with batch normalization trained on ImageNet
VGG-19 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" trained on ImageNet
Hugging Face Transformers model for image classification
Wide ResNet-101-2 model from "Wide Residual Networks" trained on ImageNet
Wide ResNet-50-2 model from "Wide Residual Networks" trained on ImageNet
YOLO-NAS is an open-source training library for advanced computer vision models. It specializes in accuracy and efficiency, supporting tasks like object detection
YOLOv2 model from "YOLO9000: Better, Faster, Stronger" trained on COCO
YOLO11-L model trained on COCO
YOLO11-L Segmentation model trained on COCO
YOLO11-M model trained on COCO
YOLO11-M Segmentation model trained on COCO
YOLO11-N model trained on COCO
YOLO11-N Segmentation model trained on COCO
YOLO11-S model trained on COCO
YOLO11-S Segmentation model trained on COCO
YOLO11-X model trained on COCO
YOLO11-X Segmentation model trained on COCO
YOLOE11-L Segmentation model
YOLOE11-M Segmentation model
YOLOE11-S Segmentation model
YOLOEv8l Segmentation model
YOLOEv8m Segmentation model
YOLOEv8s Segmentation model
YOLOv10-L model trained on COCO
YOLOv10-M model trained on COCO
YOLOv10-N model trained on COCO
YOLOv10-S model trained on COCO
YOLOv10-X model trained on COCO
Ultralytics YOLOv5l model trained on COCO
Ultralytics YOLOv5m model trained on COCO
Ultralytics YOLOv5n model trained on COCO
Ultralytics YOLOv5s model trained on COCO
Ultralytics YOLOv5x model trained on COCO
Ultralytics YOLOv8l model trained on COCO
YOLOv8l Oriented Bounding Box model
Ultralytics YOLOv8l model trained Open Images v7
Ultralytics YOLOv8l Segmentation model trained on COCO
YOLOv8l-World model
Ultralytics YOLOv8m model trained on COCO
YOLOv8m Oriented Bounding Box model
Ultralytics YOLOv8m model trained Open Images v7
Ultralytics YOLOv8m Segmentation model trained on COCO
YOLOv8m-World model
Ultralytics YOLOv8n model trained on COCO
YOLOv8n Oriented Bounding Box model
Ultralytics YOLOv8n model trained on Open Images v7
Ultralytics YOLOv8n Segmentation model trained on COCO
Ultralytics YOLOv8s model trained on COCO
YOLOv8s Oriented Bounding Box model
Ultralytics YOLOv8s model trained on Open Images v7
Ultralytics YOLOv8s Segmentation model trained on COCO
YOLOv8s-World model
Ultralytics YOLOv8x model trained on COCO
YOLOv8x Oriented Bounding Box model
Ultralytics YOLOv8x model trained Open Images v7
Ultralytics YOLOv8x Segmentation model trained on COCO
YOLOv8x-World model
YOLOv9-C model trained on COCO
YOLOv9-C Segmentation model trained on COCO
YOLOv9-E model trained on COCO
YOLOv9-E Segmentation model trained on COCO
Hugging Face Transformers model for zero-shot image classification
Hugging Face Transformers model for zero-shot object detection
Torch models#
alexnet-imagenet-torch#
AlexNet model architecture from One weird trick for parallelizing convolutional neural networks trained on ImageNet.
Details
Model name:
alexnet-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Alex Krizhevsky
Model license: BSD 3-Clause
Model size: 233.10 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, alexnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("alexnet-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
classification-transformer-torch#
Hugging Face Transformers model for image classification.
Details
Model name:
classification-transformer-torchModel source: https://huggingface.co/docs/transformers/tasks/image_classification
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? yes
Tags:
classification, logits, embeddings, torch, transformers
Requirements
Packages:
torch, torchvision, transformersCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("classification-transformer-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
clip-vit-base32-torch#
CLIP text/image encoder from Learning Transferable Visual Models From Natural Language Supervision trained on 400M text-image pairs.
Details
Model name:
clip-vit-base32-torchModel source: openai/CLIP
Model author: Alec Radford, et al.
Model license: MIT
Model size: 337.58 MB
Exposes embeddings? yes
Tags:
classification, logits, embeddings, torch, clip, zero-shot
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("clip-vit-base32-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "clip-vit-base32-torch",
24 text_prompt="A photo of a",
25 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
26)
27
28dataset.apply_model(model, label_field="predictions")
29session.refresh()
deeplabv3-resnet101-coco-torch#
DeepLabV3 model from Rethinking Atrous Convolution for Semantic Image Segmentation with ResNet-101 backbone trained on COCO.
Details
Model name:
deeplabv3-resnet101-coco-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Liang-Chieh Chen, et al.
Model license: BSD 3-Clause
Model size: 233.22 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, resnet, deeplabv3
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("deeplabv3-resnet101-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
deeplabv3-resnet50-coco-torch#
DeepLabV3 model from Rethinking Atrous Convolution for Semantic Image Segmentation with ResNet-50 backbone trained on COCO.
Details
Model name:
deeplabv3-resnet50-coco-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Liang-Chieh Chen, et al.
Model license: BSD 3-Clause
Model size: 160.51 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, resnet, deeplabv3
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("deeplabv3-resnet50-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
densenet121-imagenet-torch#
Densenet-121 model from Densely Connected Convolutional Networks trained on ImageNet.
Details
Model name:
densenet121-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Gao Huang, et al.
Model license: BSD 3-Clause
Model size: 30.84 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, densenet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("densenet121-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
densenet161-imagenet-torch#
Densenet-161 model from Densely Connected Convolutional Networks trained on ImageNet.
Details
Model name:
densenet161-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Gao Huang, et al.
Model license: BSD 3-Clause
Model size: 110.37 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, densenet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("densenet161-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
densenet169-imagenet-torch#
Densenet-169 model from Densely Connected Convolutional Networks trained on ImageNet.
Details
Model name:
densenet169-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Gao Huang, et al.
Model license: BSD 3-Clause
Model size: 54.71 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, densenet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("densenet169-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
densenet201-imagenet-torch#
Densenet-201 model from Densely Connected Convolutional Networks trained on ImageNet.
Details
Model name:
densenet201-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Gao Huang, et al.
Model license: BSD 3-Clause
Model size: 77.37 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, densenet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("densenet201-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
depth-estimation-transformer-torch#
Hugging Face Transformers model for monocular depth estimation.
Details
Model name:
depth-estimation-transformer-torchModel source: https://huggingface.co/docs/transformers/tasks/monocular_depth_estimation
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? no
Tags:
depth, torch, transformers
Requirements
Packages:
torch, torchvision, transformersCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("depth-estimation-transformer-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
detection-transformer-torch#
Hugging Face Transformers model for object detection.
Details
Model name:
detection-transformer-torchModel source: https://huggingface.co/docs/transformers/tasks/object_detection
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? yes
Tags:
detection, logits, embeddings, torch, transformers
Requirements
Packages:
torch, torchvision, transformersCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("detection-transformer-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
dinov2-vitb14-reg-torch#
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-B/14 distilled.
Details
Model name:
dinov2-vitb14-reg-torchModel source: facebookresearch/dinov2
Model author: Maxime Oquab, et al.
Model license: Apache 2.0
Model size: 330.35 MB
Exposes embeddings? yes
Tags:
embeddings, torch, dinov2
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitb14-reg-torch")
13
14embeddings = dataset.compute_embeddings(model)
dinov2-vitb14-torch#
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-B/14 distilled.
Details
Model name:
dinov2-vitb14-torchModel source: facebookresearch/dinov2
Model author: Maxime Oquab, et al.
Model license: Apache 2.0
Model size: 330.33 MB
Exposes embeddings? yes
Tags:
embeddings, torch, dinov2
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitb14-torch")
13
14embeddings = dataset.compute_embeddings(model)
dinov2-vitg14-reg-torch#
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-g/14.
Details
Model name:
dinov2-vitg14-reg-torchModel source: facebookresearch/dinov2
Model author: Maxime Oquab, et al.
Model license: Apache 2.0
Model size: 4.23 GB
Exposes embeddings? yes
Tags:
embeddings, torch, dinov2
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitg14-reg-torch")
13
14embeddings = dataset.compute_embeddings(model)
dinov2-vitg14-torch#
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-g/14.
Details
Model name:
dinov2-vitg14-torchModel source: facebookresearch/dinov2
Model author: Maxime Oquab, et al.
Model license: Apache 2.0
Model size: 4.23 GB
Exposes embeddings? yes
Tags:
embeddings, torch, dinov2
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitg14-torch")
13
14embeddings = dataset.compute_embeddings(model)
dinov2-vitl14-reg-torch#
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-L/14 distilled.
Details
Model name:
dinov2-vitl14-reg-torchModel source: facebookresearch/dinov2
Model author: Maxime Oquab, et al.
Model license: Apache 2.0
Model size: 1.13 GB
Exposes embeddings? yes
Tags:
embeddings, torch, dinov2
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitl14-reg-torch")
13
14embeddings = dataset.compute_embeddings(model)
dinov2-vitl14-torch#
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-L/14 distilled.
Details
Model name:
dinov2-vitl14-torchModel source: facebookresearch/dinov2
Model author: Maxime Oquab, et al.
Model license: Apache 2.0
Model size: 1.13 GB
Exposes embeddings? yes
Tags:
embeddings, torch, dinov2
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitl14-torch")
13
14embeddings = dataset.compute_embeddings(model)
dinov2-vits14-reg-torch#
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-S/14 distilled.
Details
Model name:
dinov2-vits14-reg-torchModel source: facebookresearch/dinov2
Model author: Maxime Oquab, et al.
Model license: Apache 2.0
Model size: 84.20 MB
Exposes embeddings? yes
Tags:
embeddings, torch, dinov2
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vits14-reg-torch")
13
14embeddings = dataset.compute_embeddings(model)
dinov2-vits14-torch#
DINOv2: Learning Robust Visual Features without Supervision. Model: ViT-S/14 distilled.
Details
Model name:
dinov2-vits14-torchModel source: facebookresearch/dinov2
Model author: Maxime Oquab, et al.
Model license: Apache 2.0
Model size: 84.19 MB
Exposes embeddings? yes
Tags:
embeddings, torch, dinov2
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vits14-torch")
13
14embeddings = dataset.compute_embeddings(model)
faster-rcnn-resnet50-fpn-coco-torch#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with ResNet-50 FPN backbone trained on COCO.
Details
Model name:
faster-rcnn-resnet50-fpn-coco-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Shaoqing Ren, et al.
Model license: BSD 3-Clause
Model size: 159.74 MB
Exposes embeddings? no
Tags:
detection, coco, torch, faster-rcnn, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet50-fpn-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
fcn-resnet101-coco-torch#
FCN model from Fully Convolutional Networks for Semantic Segmentation with ResNet-101 backbone trained on COCO.
Details
Model name:
fcn-resnet101-coco-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Jonathan Long, et al.
Model license: BSD 3-Clause
Model size: 207.71 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, fcn, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("fcn-resnet101-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
fcn-resnet50-coco-torch#
FCN model from Fully Convolutional Networks for Semantic Segmentation with ResNet-50 backbone trained on COCO.
Details
Model name:
fcn-resnet50-coco-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Jonathan Long, et al.
Model license: BSD 3-Clause
Model size: 135.01 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, fcn, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("fcn-resnet50-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
googlenet-imagenet-torch#
GoogLeNet (Inception v1) model from Going Deeper with Convolutions trained on ImageNet.
Details
Model name:
googlenet-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Christian Szegedy, et al.
Model license: BSD 3-Clause
Model size: 49.73 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, googlenet
Requirements
Packages:
scipy, torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("googlenet-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
group-vit-segmentation-transformer-torch#
Hugging Face Transformers model for zero-shot semantic segmentation.
Details
Model name:
group-vit-segmentation-transformer-torchModel source: https://huggingface.co/docs/transformers/en/tasks/mask_generation
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? yes
Tags:
segmentation, embeddings, torch, transformers, zero-shot
Requirements
Packages:
torch, torchvision, transformersCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("group-vit-segmentation-transformer-torch",
13 text_prompt="A photo of a",
14 classes=["person", "dog", "cat", "bird", "car", "tree", "other"])
15
16dataset.apply_model(model, label_field="predictions")
17
18session = fo.launch_app(dataset)
inception-v3-imagenet-torch#
Inception v3 model from Rethinking the Inception Architecture for Computer Vision trained on ImageNet.
Details
Model name:
inception-v3-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Christian Szegedy, et al.
Model license: BSD 3-Clause
Model size: 103.81 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, inception
Requirements
Packages:
scipy, torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("inception-v3-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
keypoint-rcnn-resnet50-fpn-coco-torch#
Keypoint R-CNN model from Mask R-CNN with ResNet-50 FPN backbone trained on COCO.
Details
Model name:
keypoint-rcnn-resnet50-fpn-coco-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Kaiming He, et al.
Model license: BSD 3-Clause
Model size: 226.05 MB
Exposes embeddings? no
Tags:
keypoints, coco, torch, keypoint-rcnn, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("keypoint-rcnn-resnet50-fpn-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
mask-rcnn-resnet50-fpn-coco-torch#
Mask R-CNN model from Mask R-CNN with ResNet-50 FPN backbone trained on COCO.
Details
Model name:
mask-rcnn-resnet50-fpn-coco-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Kaiming He, et al.
Model license: BSD 3-Clause
Model size: 169.84 MB
Exposes embeddings? no
Tags:
instances, coco, torch, mask-rcnn, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-resnet50-fpn-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
med-sam-2-video-torch#
Fine-tuned SAM2-hiera-tiny model from Medical SAM 2 - Segment Medical Images as Video via Segment Anything Model 2.
Details
Model name:
med-sam-2-video-torchModel source: MedicineToken/Medical-SAM2
Model author: Jiayuan Zhu, et al.
Model license: Apache 2.0
Model size: 74.46 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot, video, med-SAM
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3from fiftyone import ViewField as F
4from fiftyone.utils.huggingface import load_from_hub
5
6dataset = load_from_hub("Voxel51/BTCV-CT-as-video-MedSAM2-dataset")[:2]
7
8# Retaining detections from a single frame in the middle
9# Note that SAM2 only propagates segmentation masks forward in a video
10(
11 dataset
12 .match_frames(F("frame_number") != 100)
13 .set_field("frames.gt_detections", None)
14 .save()
15)
16
17model = foz.load_zoo_model("med-sam-2-video-torch")
18
19# Segment inside boxes and propagate to all frames
20dataset.apply_model(
21 model,
22 label_field="pred_segmentations",
23 prompt_field="frames.gt_detections",
24)
25
26session = fo.launch_app(dataset)
mnasnet0.5-imagenet-torch#
MNASNet model from MnasNet: Platform-Aware Neural Architecture Search for Mobile with depth multiplier of 0.5 trained on ImageNet.
Details
Model name:
mnasnet0.5-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Mingxing Tan, et al.
Model license: BSD 3-Clause
Model size: 8.59 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, mnasnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("mnasnet0.5-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
mnasnet1.0-imagenet-torch#
MNASNet model from MnasNet: Platform-Aware Neural Architecture Search for Mobile with depth multiplier of 1.0 trained on ImageNet.
Details
Model name:
mnasnet1.0-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Mingxing Tan, et al.
Model license: BSD 3-Clause
Model size: 16.92 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, mnasnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("mnasnet1.0-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
mobilenet-v2-imagenet-torch#
MobileNetV2 model from MobileNetV2: Inverted Residuals and Linear Bottlenecks trained on ImageNet.
Details
Model name:
mobilenet-v2-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Mark Sandler, et al.
Model license: BSD 3-Clause
Model size: 13.55 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, mobilenet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("mobilenet-v2-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
omdet-turbo-swin-tiny-torch#
Hugging Face Transformers OmDet-Turbo.
Details
Model name:
omdet-turbo-swin-tiny-torchModel source: https://huggingface.co/docs/transformers/tasks/zero_shot_object_detection
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? yes
Tags:
detection, logits, embeddings, torch, transformers, zero-shot
Requirements
Packages:
torch, torchvision, transformers>=4.51CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12
13classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
14
15model = foz.load_zoo_model(
16 "omdet-turbo-swin-tiny-torch",
17 classes=classes,
18)
19
20dataset.apply_model(model, label_field="predictions")
21
22session = fo.launch_app(dataset)
open-clip-torch#
OPEN CLIP text/image encoder from Learning Transferable Visual Models From Natural Language Supervision trained on 400M text-image pairs.
Details
Model name:
open-clip-torchModel source: mlfoundations/open_clip
Model author: Gabriel Ilharco, et al.
Model license: MIT
Exposes embeddings? yes
Tags:
classification, logits, embeddings, torch, clip, zero-shot
Requirements
Packages:
torch, torchvision, open_clip_torchCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("open-clip-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "open-clip-torch",
24 text_prompt="A photo of a",
25 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
26)
27
28dataset.apply_model(model, label_field="predictions")
29session.refresh()
owlvit-base-patch16-torch#
Hugging Face Transformers OWL-ViT.
Details
Model name:
owlvit-base-patch16-torchModel source: https://huggingface.co/docs/transformers/tasks/zero_shot_object_detection
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? yes
Tags:
detection, logits, embeddings, torch, transformers, zero-shot
Requirements
Packages:
torch, torchvision, transformersCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12
13classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
14
15model = foz.load_zoo_model(
16 "owlvit-base-patch16-torch",
17 classes=classes,
18)
19
20dataset.apply_model(model, label_field="predictions")
21
22session = fo.launch_app(dataset)
resnet101-imagenet-torch#
ResNet-101 model from Deep Residual Learning for Image Recognition trained on ImageNet.
Details
Model name:
resnet101-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Kaiming He, et al.
Model license: BSD 3-Clause
Model size: 170.45 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("resnet101-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
resnet152-imagenet-torch#
ResNet-152 model from Deep Residual Learning for Image Recognition trained on ImageNet.
Details
Model name:
resnet152-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Kaiming He, et al.
Model license: BSD 3-Clause
Model size: 230.34 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("resnet152-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
resnet18-imagenet-torch#
ResNet-18 model from Deep Residual Learning for Image Recognition trained on ImageNet.
Details
Model name:
resnet18-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Kaiming He, et al.
Model license: BSD 3-Clause
Model size: 44.66 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("resnet18-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
resnet34-imagenet-torch#
ResNet-34 model from Deep Residual Learning for Image Recognition trained on ImageNet.
Details
Model name:
resnet34-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Kaiming He, et al.
Model license: BSD 3-Clause
Model size: 83.26 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("resnet34-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
resnet50-imagenet-torch#
ResNet-50 model from Deep Residual Learning for Image Recognition trained on ImageNet.
Details
Model name:
resnet50-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Kaiming He, et al.
Model license: BSD 3-Clause
Model size: 97.75 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("resnet50-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
resnext101-32x8d-imagenet-torch#
ResNeXt-101 32x8d model from Aggregated Residual Transformations for Deep Neural Networks trained on ImageNet.
Details
Model name:
resnext101-32x8d-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Saining Xie, et al.
Model license: BSD 3-Clause
Model size: 339.59 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, resnext
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("resnext101-32x8d-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
resnext50-32x4d-imagenet-torch#
ResNeXt-50 32x4d model from Aggregated Residual Transformations for Deep Neural Networks trained on ImageNet.
Details
Model name:
resnext50-32x4d-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Saining Xie, et al.
Model license: BSD 3-Clause
Model size: 95.79 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, resnext
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("resnext50-32x4d-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
retinanet-resnet50-fpn-coco-torch#
RetinaNet model from Focal Loss for Dense Object Detection with ResNet-50 FPN backbone trained on COCO.
Details
Model name:
retinanet-resnet50-fpn-coco-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Tsung-Yi Lin, et al.
Model license: BSD 3-Clause
Model size: 130.27 MB
Exposes embeddings? no
Tags:
detection, coco, torch, retinanet, resnet
Requirements
Packages:
torch, torchvision>=0.8.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("retinanet-resnet50-fpn-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
rtdetr-l-coco-torch#
RT-DETR-l model trained on COCO.
Details
Model name:
rtdetr-l-coco-torchModel source: https://docs.ultralytics.com/models/rtdetr/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 63.43 MB
Exposes embeddings? no
Tags:
detection, coco, torch, transformer, rtdetr
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("rtdetr-l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
rtdetr-x-coco-torch#
RT-DETR-x model trained on COCO.
Details
Model name:
rtdetr-x-coco-torchModel source: https://docs.ultralytics.com/models/rtdetr/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 129.47 MB
Exposes embeddings? no
Tags:
detection, coco, torch, transformer, rtdetr
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("rtdetr-x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
segment-anything-2-hiera-base-plus-image-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2-hiera-base-plus-image-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 308.51 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2-hiera-base-plus-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-2-hiera-base-plus-video-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2-hiera-base-plus-video-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 308.51 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot, video
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3from fiftyone import ViewField as F
4
5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
6
7# Only retain detections in the first frame
8(
9 dataset
10 .match_frames(F("frame_number") > 1)
11 .set_field("frames.detections", None)
12 .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2-hiera-base-plus-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19 model,
20 label_field="segmentations",
21 prompt_field="frames.detections", # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)
segment-anything-2-hiera-large-image-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2-hiera-large-image-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 856.35 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2-hiera-large-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-2-hiera-large-video-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2-hiera-large-video-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 856.35 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot, video
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3from fiftyone import ViewField as F
4
5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
6
7# Only retain detections in the first frame
8(
9 dataset
10 .match_frames(F("frame_number") > 1)
11 .set_field("frames.detections", None)
12 .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2-hiera-large-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19 model,
20 label_field="segmentations",
21 prompt_field="frames.detections", # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)
segment-anything-2-hiera-small-image-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2-hiera-small-image-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 175.77 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2-hiera-small-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-2-hiera-small-video-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2-hiera-small-video-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 175.77 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot, video
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3from fiftyone import ViewField as F
4
5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
6
7# Only retain detections in the first frame
8(
9 dataset
10 .match_frames(F("frame_number") > 1)
11 .set_field("frames.detections", None)
12 .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2-hiera-small-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19 model,
20 label_field="segmentations",
21 prompt_field="frames.detections", # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)
segment-anything-2-hiera-tiny-image-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2-hiera-tiny-image-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 148.68 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2-hiera-tiny-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-2-hiera-tiny-video-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2-hiera-tiny-video-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 148.68 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot, video
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3from fiftyone import ViewField as F
4
5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
6
7# Only retain detections in the first frame
8(
9 dataset
10 .match_frames(F("frame_number") > 1)
11 .set_field("frames.detections", None)
12 .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2-hiera-tiny-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19 model,
20 label_field="segmentations",
21 prompt_field="frames.detections", # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)
segment-anything-2.1-hiera-base-plus-image-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2.1-hiera-base-plus-image-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 308.62 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2.1-hiera-base-plus-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-2.1-hiera-base-plus-video-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2.1-hiera-base-plus-video-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 308.62 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot, video
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3from fiftyone import ViewField as F
4
5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
6
7# Only retain detections in the first frame
8(
9 dataset
10 .match_frames(F("frame_number") > 1)
11 .set_field("frames.detections", None)
12 .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2.1-hiera-base-plus-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19 model,
20 label_field="segmentations",
21 prompt_field="frames.detections", # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)
segment-anything-2.1-hiera-large-image-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2.1-hiera-large-image-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 856.48 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2.1-hiera-large-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-2.1-hiera-large-video-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2.1-hiera-large-video-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 856.48 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot, video
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3from fiftyone import ViewField as F
4
5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
6
7# Only retain detections in the first frame
8(
9 dataset
10 .match_frames(F("frame_number") > 1)
11 .set_field("frames.detections", None)
12 .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2.1-hiera-large-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19 model,
20 label_field="segmentations",
21 prompt_field="frames.detections", # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)
segment-anything-2.1-hiera-small-image-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2.1-hiera-small-image-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 175.87 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2.1-hiera-small-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-2.1-hiera-small-video-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2.1-hiera-small-video-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 175.87 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot, video
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3from fiftyone import ViewField as F
4
5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
6
7# Only retain detections in the first frame
8(
9 dataset
10 .match_frames(F("frame_number") > 1)
11 .set_field("frames.detections", None)
12 .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2.1-hiera-small-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19 model,
20 label_field="segmentations",
21 prompt_field="frames.detections", # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)
segment-anything-2.1-hiera-tiny-image-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2.1-hiera-tiny-image-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 148.68 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2.1-hiera-tiny-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-2.1-hiera-tiny-video-torch#
Segment Anything Model 2 (SAM2) from SAM2: Segment Anything in Images and Videos.
Details
Model name:
segment-anything-2.1-hiera-tiny-video-torchModel source: https://ai.meta.com/sam2/
Model author: Nikhila Ravi, et al.
Model license: Apache 2.0,BSD 3-Clause
Model size: 148.68 MB
Exposes embeddings? no
Tags:
segment-anything, torch, zero-shot, video
Requirements
Packages:
torch, torchvision, sam2CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3from fiftyone import ViewField as F
4
5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
6
7# Only retain detections in the first frame
8(
9 dataset
10 .match_frames(F("frame_number") > 1)
11 .set_field("frames.detections", None)
12 .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2.1-hiera-tiny-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19 model,
20 label_field="segmentations",
21 prompt_field="frames.detections", # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)
segment-anything-vitb-torch#
Segment Anything Model (SAM) from Segment Anything with ViT-B/16 backbone trained on SA-1B.
Details
Model name:
segment-anything-vitb-torchModel source: https://segment-anything.com
Model author: Alexander Kirillov, et al.
Model license: Apache 2.0
Model size: 357.67 MB
Exposes embeddings? no
Tags:
segment-anything, sa-1b, torch, zero-shot
Requirements
Packages:
torch, torchvision, segment-anythingCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-vitb-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-vith-torch#
Segment Anything Model (SAM) from Segment Anything with ViT-H/16 backbone trained on SA-1B.
Details
Model name:
segment-anything-vith-torchModel source: https://segment-anything.com
Model author: Alexander Kirillov, et al.
Model license: Apache 2.0
Model size: 2.39 GB
Exposes embeddings? no
Tags:
segment-anything, sa-1b, torch, zero-shot
Requirements
Packages:
torch, torchvision, segment-anythingCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-vith-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segment-anything-vitl-torch#
Segment Anything Model (SAM) from Segment Anything with ViT-L/16 backbone trained on SA-1B.
Details
Model name:
segment-anything-vitl-torchModel source: https://segment-anything.com
Model author: Alexander Kirillov, et al.
Model license: Apache 2.0
Model size: 1.16 GB
Exposes embeddings? no
Tags:
segment-anything, sa-1b, torch, zero-shot
Requirements
Packages:
torch, torchvision, segment-anythingCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-vitl-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16 model,
17 label_field="segmentations",
18 prompt_field="ground_truth", # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)
segmentation-transformer-torch#
Hugging Face Transformers model for semantic segmentation.
Details
Model name:
segmentation-transformer-torchModel source: https://huggingface.co/docs/transformers/tasks/semantic_segmentation
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? no
Tags:
segmentation, torch, transformers
Requirements
Packages:
torch, torchvision, transformersCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("segmentation-transformer-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
shufflenetv2-0.5x-imagenet-torch#
ShuffleNetV2 model from ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design with 0.5x output channels trained on ImageNet.
Details
Model name:
shufflenetv2-0.5x-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Ningning Ma, et al.
Model license: BSD 3-Clause
Model size: 5.28 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, shufflenet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("shufflenetv2-0.5x-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
shufflenetv2-1.0x-imagenet-torch#
ShuffleNetV2 model from ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design with 1.0x output channels trained on ImageNet.
Details
Model name:
shufflenetv2-1.0x-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Ningning Ma, et al.
Model license: BSD 3-Clause
Model size: 8.79 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, shufflenet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("shufflenetv2-1.0x-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
siglip-base-patch16-224-torch#
Hugging Face Transformers model for zero-shot image classification.
Details
Model name:
siglip-base-patch16-224-torchModel source: https://huggingface.co/docs/transformers/tasks/zero_shot_image_classification
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? yes
Tags:
classification, logits, embeddings, torch, transformers, zero-shot
Requirements
Packages:
torch, torchvision, transformers>=4.51CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12
13classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
14
15model = foz.load_zoo_model(
16 "siglip-base-patch16-224-torch",
17 classes=classes,
18)
19
20dataset.apply_model(model, label_field="predictions")
21
22session = fo.launch_app(dataset)
squeezenet-1.1-imagenet-torch#
SqueezeNet 1.1 model from the official SqueezeNet repo trained on ImageNet.
Details
Model name:
squeezenet-1.1-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Forrest Iandola
Model license: BSD 2-Clause
Model size: 4.74 MB
Exposes embeddings? no
Tags:
classification, imagenet, torch, squeezenet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("squeezenet-1.1-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
squeezenet-imagenet-torch#
SqueezeNet model from SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size trained on ImageNet.
Details
Model name:
squeezenet-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Forrest Iandola
Model license: BSD 2-Clause
Model size: 4.79 MB
Exposes embeddings? no
Tags:
classification, imagenet, torch, squeezenet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("squeezenet-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
vgg11-bn-imagenet-torch#
VGG-11 model from Very Deep Convolutional Networks for Large-Scale Image Recognition with batch normalization trained on ImageNet.
Details
Model name:
vgg11-bn-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Karen Simonyan, et al.
Model license: BSD 3-Clause
Model size: 506.88 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, vgg
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vgg11-bn-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
vgg11-imagenet-torch#
VGG-11 model from Very Deep Convolutional Networks for Large-Scale Image Recognition trained on ImageNet.
Details
Model name:
vgg11-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Karen Simonyan, et al.
Model license: BSD 3-Clause
Model size: 506.84 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, vgg
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vgg11-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
vgg13-bn-imagenet-torch#
VGG-13 model from Very Deep Convolutional Networks for Large-Scale Image Recognition with batch normalization trained on ImageNet.
Details
Model name:
vgg13-bn-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Karen Simonyan, et al.
Model license: BSD 3-Clause
Model size: 507.59 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, vgg
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vgg13-bn-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
vgg13-imagenet-torch#
VGG-13 model from Very Deep Convolutional Networks for Large-Scale Image Recognition trained on ImageNet.
Details
Model name:
vgg13-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Karen Simonyan, et al.
Model license: BSD 3-Clause
Model size: 507.54 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, vgg
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vgg13-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
vgg16-bn-imagenet-torch#
VGG-16 model from Very Deep Convolutional Networks for Large-Scale Image Recognition with batch normalization trained on ImageNet.
Details
Model name:
vgg16-bn-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Karen Simonyan, et al.
Model license: BSD 3-Clause
Model size: 527.87 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, vgg
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vgg16-bn-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
vgg16-imagenet-torch#
VGG-16 model from Very Deep Convolutional Networks for Large-Scale Image Recognition trained on ImageNet.
Details
Model name:
vgg16-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Karen Simonyan, et al.
Model license: BSD 3-Clause
Model size: 527.80 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, vgg
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vgg16-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
vgg19-bn-imagenet-torch#
VGG-19 model from Very Deep Convolutional Networks for Large-Scale Image Recognition with batch normalization trained on ImageNet.
Details
Model name:
vgg19-bn-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Karen Simonyan, et al.
Model license: BSD 3-Clause
Model size: 548.14 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, vgg
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vgg19-bn-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
vgg19-imagenet-torch#
VGG-19 model from Very Deep Convolutional Networks for Large-Scale Image Recognition trained on ImageNet.
Details
Model name:
vgg19-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Karen Simonyan, et al.
Model license: BSD 3-Clause
Model size: 548.05 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, vgg
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vgg19-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
vit-base-patch16-224-imagenet-torch#
Hugging Face Transformers model for image classification.
Details
Model name:
vit-base-patch16-224-imagenet-torchModel source: https://huggingface.co/docs/transformers/tasks/image_classification
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? yes
Tags:
classification, logits, embeddings, torch, transformers
Requirements
Packages:
torch, torchvision, transformersCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vit-base-patch16-224-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
wide-resnet101-2-imagenet-torch#
Wide ResNet-101-2 model from Wide Residual Networks trained on ImageNet.
Details
Model name:
wide-resnet101-2-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Sergey Zagoruyko, et al.
Model license: BSD 3-Clause
Model size: 242.90 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, wide-resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("wide-resnet101-2-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
wide-resnet50-2-imagenet-torch#
Wide ResNet-50-2 model from Wide Residual Networks trained on ImageNet.
Details
Model name:
wide-resnet50-2-imagenet-torchModel source: https://pytorch.org/vision/main/models.html
Model author: Sergey Zagoruyko, et al.
Model license: BSD 3-Clause
Model size: 131.82 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, torch, wide-resnet
Requirements
Packages:
torch, torchvisionCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("wide-resnet50-2-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
yolo-nas-torch#
YOLO-NAS is an open-source training library for advanced computer vision models. It specializes in accuracy and efficiency, supporting tasks like object detection.
Details
Model name:
yolo-nas-torchModel source: Deci-AI/super-gradients
Model author: Shay Aharon, et al.
Model license: Apache 2.0
Exposes embeddings? no
Tags:
detection, torch, yolo
Requirements
Packages:
torch, torchvision, super-gradientsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo-nas-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11l-coco-torch#
YOLO11-L model trained on COCO.
Details
Model name:
yolo11l-coco-torchModel source: https://docs.ultralytics.com/models/yolov11/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 49.01 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11l-seg-coco-torch#
YOLO11-L Segmentation model trained on COCO.
Details
Model name:
yolo11l-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolo11/#__tabbed_1_2
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 53.50 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11l-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11m-coco-torch#
YOLO11-M model trained on COCO.
Details
Model name:
yolo11m-coco-torchModel source: https://docs.ultralytics.com/models/yolov11/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 38.80 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11m-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11m-seg-coco-torch#
YOLO11-M Segmentation model trained on COCO.
Details
Model name:
yolo11m-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolo11/#__tabbed_1_2
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 43.30 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11m-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11n-coco-torch#
YOLO11-N model trained on COCO.
Details
Model name:
yolo11n-coco-torchModel source: https://docs.ultralytics.com/models/yolov11/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 5.35 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11n-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11n-seg-coco-torch#
YOLO11-N Segmentation model trained on COCO.
Details
Model name:
yolo11n-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolo11/#__tabbed_1_2
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 5.90 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11n-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11s-coco-torch#
YOLO11-S model trained on COCO.
Details
Model name:
yolo11s-coco-torchModel source: https://docs.ultralytics.com/models/yolov11/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 18.42 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11s-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11s-seg-coco-torch#
YOLO11-S Segmentation model trained on COCO.
Details
Model name:
yolo11s-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolo11/#__tabbed_1_2
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 19.71 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11s-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11x-coco-torch#
YOLO11-X model trained on COCO.
Details
Model name:
yolo11x-coco-torchModel source: https://docs.ultralytics.com/models/yolov11/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 109.33 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolo11x-seg-coco-torch#
YOLO11-X Segmentation model trained on COCO.
Details
Model name:
yolo11x-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolo11/#__tabbed_1_2
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 119.30 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11x-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yoloe11l-seg-torch#
YOLOE11-L Segmentation model.
Details
Model name:
yoloe11l-seg-torchModel source: https://docs.ultralytics.com/models/yoloe
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 67.69 MB
Exposes embeddings? no
Tags:
segmentation, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloe11l-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yoloe11l-seg-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yoloe11m-seg-torch#
YOLOE11-M Segmentation model.
Details
Model name:
yoloe11m-seg-torchModel source: https://docs.ultralytics.com/models/yoloe
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 57.48 MB
Exposes embeddings? no
Tags:
segmentation, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloe11m-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yoloe11m-seg-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yoloe11s-seg-torch#
YOLOE11-S Segmentation model.
Details
Model name:
yoloe11s-seg-torchModel source: https://docs.ultralytics.com/models/yoloe
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 26.52 MB
Exposes embeddings? no
Tags:
segmentation, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloe11s-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yoloe11s-seg-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yoloev8l-seg-torch#
YOLOEv8l Segmentation model.
Details
Model name:
yoloev8l-seg-torchModel source: https://docs.ultralytics.com/models/yoloe
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 102.43 MB
Exposes embeddings? no
Tags:
segmentation, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloev8l-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yoloev8l-seg-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yoloev8m-seg-torch#
YOLOEv8m Segmentation model.
Details
Model name:
yoloev8m-seg-torchModel source: https://docs.ultralytics.com/models/yoloe
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 62.75 MB
Exposes embeddings? no
Tags:
segmentation, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloev8m-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yoloev8m-seg-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yoloev8s-seg-torch#
YOLOEv8s Segmentation model.
Details
Model name:
yoloev8s-seg-torchModel source: https://docs.ultralytics.com/models/yoloe
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 29.69 MB
Exposes embeddings? no
Tags:
segmentation, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloev8s-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yoloev8s-seg-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yolov10l-coco-torch#
YOLOv10-L model trained on COCO.
Details
Model name:
yolov10l-coco-torchModel source: https://docs.ultralytics.com/models/yolov10/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 50.00 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov10m-coco-torch#
YOLOv10-M model trained on COCO.
Details
Model name:
yolov10m-coco-torchModel source: https://docs.ultralytics.com/models/yolov10/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 32.09 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10m-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov10n-coco-torch#
YOLOv10-N model trained on COCO.
Details
Model name:
yolov10n-coco-torchModel source: https://docs.ultralytics.com/models/yolov10/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 5.59 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10n-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov10s-coco-torch#
YOLOv10-S model trained on COCO.
Details
Model name:
yolov10s-coco-torchModel source: https://docs.ultralytics.com/models/yolov10/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 15.85 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10s-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov10x-coco-torch#
YOLOv10-X model trained on COCO.
Details
Model name:
yolov10x-coco-torchModel source: https://docs.ultralytics.com/models/yolov10/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 61.41 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov5l-coco-torch#
Ultralytics YOLOv5l model trained on COCO.
Details
Model name:
yolov5l-coco-torchModel source: https://pytorch.org/hub/ultralytics_yolov5
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 101.96 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov5m-coco-torch#
Ultralytics YOLOv5m model trained on COCO.
Details
Model name:
yolov5m-coco-torchModel source: https://pytorch.org/hub/ultralytics_yolov5
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 48.25 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5m-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov5n-coco-torch#
Ultralytics YOLOv5n model trained on COCO.
Details
Model name:
yolov5n-coco-torchModel source: https://pytorch.org/hub/ultralytics_yolov5
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 5.31 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5n-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov5s-coco-torch#
Ultralytics YOLOv5s model trained on COCO.
Details
Model name:
yolov5s-coco-torchModel source: https://pytorch.org/hub/ultralytics_yolov5
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 17.72 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5s-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov5x-coco-torch#
Ultralytics YOLOv5x model trained on COCO.
Details
Model name:
yolov5x-coco-torchModel source: https://pytorch.org/hub/ultralytics_yolov5
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 186.09 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8l-coco-torch#
Ultralytics YOLOv8l model trained on COCO.
Details
Model name:
yolov8l-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 83.70 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8l-obb-dotav1-torch#
YOLOv8l Oriented Bounding Box model.
Details
Model name:
yolov8l-obb-dotav1-torchModel source: https://docs.ultralytics.com/tasks/obb/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 85.36 MB
Exposes embeddings? no
Tags:
detection, torch, yolo, polylines, obb
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8l-oiv7-torch#
Ultralytics YOLOv8l model trained Open Images v7.
Details
Model name:
yolov8l-oiv7-torchModel source: https://docs.ultralytics.com/datasets/detect/open-images-v7
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 83.70 MB
Exposes embeddings? no
Tags:
detection, oiv7, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8l-seg-coco-torch#
Ultralytics YOLOv8l Segmentation model trained on COCO.
Details
Model name:
yolov8l-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 88.11 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8l-world-torch#
YOLOv8l-World model.
Details
Model name:
yolov8l-world-torchModel source: https://docs.ultralytics.com/models/yolo-world/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 91.23 MB
Exposes embeddings? no
Tags:
detection, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-world-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yolov8l-world-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yolov8m-coco-torch#
Ultralytics YOLOv8m model trained on COCO.
Details
Model name:
yolov8m-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 49.70 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8m-obb-dotav1-torch#
YOLOv8m Oriented Bounding Box model.
Details
Model name:
yolov8m-obb-dotav1-torchModel source: https://docs.ultralytics.com/tasks/obb/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 50.84 MB
Exposes embeddings? no
Tags:
detection, torch, yolo, polylines, obb
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8m-oiv7-torch#
Ultralytics YOLOv8m model trained Open Images v7.
Details
Model name:
yolov8m-oiv7-torchModel source: https://docs.ultralytics.com/datasets/detect/open-images-v7
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 50.29 MB
Exposes embeddings? no
Tags:
detection, oiv7, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8m-seg-coco-torch#
Ultralytics YOLOv8m Segmentation model trained on COCO.
Details
Model name:
yolov8m-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 52.36 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8m-world-torch#
YOLOv8m-World model.
Details
Model name:
yolov8m-world-torchModel source: https://docs.ultralytics.com/models/yolo-world/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 55.89 MB
Exposes embeddings? no
Tags:
detection, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-world-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yolov8m-world-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yolov8n-coco-torch#
Ultralytics YOLOv8n model trained on COCO.
Details
Model name:
yolov8n-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 6.23 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8n-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8n-obb-dotav1-torch#
YOLOv8n Oriented Bounding Box model.
Details
Model name:
yolov8n-obb-dotav1-torchModel source: https://docs.ultralytics.com/tasks/obb/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 6.24 MB
Exposes embeddings? no
Tags:
detection, torch, yolo, polylines, obb
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8n-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8n-oiv7-torch#
Ultralytics YOLOv8n model trained on Open Images v7.
Details
Model name:
yolov8n-oiv7-torchModel source: https://docs.ultralytics.com/datasets/detect/open-images-v7
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 6.89 MB
Exposes embeddings? no
Tags:
detection, oiv7, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8n-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8n-seg-coco-torch#
Ultralytics YOLOv8n Segmentation model trained on COCO.
Details
Model name:
yolov8n-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 6.73 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8n-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8s-coco-torch#
Ultralytics YOLOv8s model trained on COCO.
Details
Model name:
yolov8s-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 21.53 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8s-obb-dotav1-torch#
YOLOv8s Oriented Bounding Box model.
Details
Model name:
yolov8s-obb-dotav1-torchModel source: https://docs.ultralytics.com/tasks/obb/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 22.17 MB
Exposes embeddings? no
Tags:
detection, torch, yolo, polylines, obb
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8s-oiv7-torch#
Ultralytics YOLOv8s model trained on Open Images v7.
Details
Model name:
yolov8s-oiv7-torchModel source: https://docs.ultralytics.com/datasets/detect/open-images-v7
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 21.92 MB
Exposes embeddings? no
Tags:
detection, oiv7, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8s-seg-coco-torch#
Ultralytics YOLOv8s Segmentation model trained on COCO.
Details
Model name:
yolov8s-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 22.79 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8s-world-torch#
YOLOv8s-World model.
Details
Model name:
yolov8s-world-torchModel source: https://docs.ultralytics.com/models/yolo-world/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 25.91 MB
Exposes embeddings? no
Tags:
detection, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-world-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yolov8s-world-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yolov8x-coco-torch#
Ultralytics YOLOv8x model trained on COCO.
Details
Model name:
yolov8x-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 130.53 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8x-obb-dotav1-torch#
YOLOv8x Oriented Bounding Box model.
Details
Model name:
yolov8x-obb-dotav1-torchModel source: https://docs.ultralytics.com/tasks/obb/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 133.07 MB
Exposes embeddings? no
Tags:
detection, torch, yolo, polylines, obb
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8x-oiv7-torch#
Ultralytics YOLOv8x model trained Open Images v7.
Details
Model name:
yolov8x-oiv7-torchModel source: https://docs.ultralytics.com/datasets/detect/open-images-v7
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 130.53 MB
Exposes embeddings? no
Tags:
detection, oiv7, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8x-seg-coco-torch#
Ultralytics YOLOv8x Segmentation model trained on COCO.
Details
Model name:
yolov8x-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolov8/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 137.40 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralyticsCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov8x-world-torch#
YOLOv8x-World model.
Details
Model name:
yolov8x-world-torchModel source: https://docs.ultralytics.com/models/yolo-world/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 141.11 MB
Exposes embeddings? no
Tags:
detection, torch, yolo, zero-shot
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-world-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23 "yolov8x-world-torch",
24 classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()
yolov9c-coco-torch#
YOLOv9-C model trained on COCO.
Details
Model name:
yolov9c-coco-torchModel source: https://docs.ultralytics.com/models/yolov9/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 49.40 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov9c-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov9c-seg-coco-torch#
YOLOv9-C Segmentation model trained on COCO.
Details
Model name:
yolov9c-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolov9/#__tabbed_1_2
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 53.86 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.42CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov9c-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov9e-coco-torch#
YOLOv9-E model trained on COCO.
Details
Model name:
yolov9e-coco-torchModel source: https://docs.ultralytics.com/models/yolov9/
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 112.09 MB
Exposes embeddings? no
Tags:
detection, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov9e-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
yolov9e-seg-coco-torch#
YOLOv9-E Segmentation model trained on COCO.
Details
Model name:
yolov9e-seg-coco-torchModel source: https://docs.ultralytics.com/models/yolov9/#__tabbed_1_2
Model author: Glenn Jocher, et al.
Model license: AGPL-3.0
Model size: 116.55 MB
Exposes embeddings? no
Tags:
segmentation, coco, torch, yolo
Requirements
Packages:
torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.42CPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov9e-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
zero-shot-classification-transformer-torch#
Hugging Face Transformers model for zero-shot image classification.
Details
Model name:
zero-shot-classification-transformer-torchModel source: https://huggingface.co/docs/transformers/tasks/zero_shot_image_classification
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? yes
Tags:
classification, logits, embeddings, torch, transformers, zero-shot
Requirements
Packages:
torch, torchvision, transformersCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
13
14model = foz.load_zoo_model(
15 "zero-shot-classification-transformer-torch",
16 classes=classes,
17)
18
19dataset.apply_model(model, label_field="predictions")
20
21session = fo.launch_app(dataset)
22
23# some models make require additional arguments
24# check the Hugging Face docs to see if any are needed
25
26# for example, AltCLIP requires `padding=True` in its processor
27model = foz.load_zoo_model(
28 "zero-shot-classification-transformer-torch",
29 classes=classes,
30 name_or_path="BAAI/AltCLIP",
31 transformers_processor_kwargs={
32 "padding": True,
33 }
34)
35
36dataset.apply_model(model, label_field="predictions")
37
38session = fo.launch_app(dataset)
zero-shot-detection-transformer-torch#
Hugging Face Transformers model for zero-shot object detection.
Details
Model name:
zero-shot-detection-transformer-torchModel source: https://huggingface.co/docs/transformers/tasks/zero_shot_object_detection
Model author: Thomas Wolf, et al.
Model license: Apache 2.0
Exposes embeddings? yes
Tags:
detection, logits, embeddings, torch, transformers, zero-shot
Requirements
Packages:
torch, torchvision, transformersCPU support
yes
GPU support
yes
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
13
14model = foz.load_zoo_model(
15 "zero-shot-detection-transformer-torch",
16 classes=classes,
17)
18
19dataset.apply_model(model, label_field="predictions")
20
21session = fo.launch_app(dataset)
TensorFlow models#
centernet-hg104-1024-coco-tf2#
CenterNet model from Objects as Points with the Hourglass-104 backbone trained on COCO resized to 1024x1024.
Details
Model name:
centernet-hg104-1024-coco-tf2Model source: tensorflow/models
Model author: Xingyi Zhou, et al.
Model license: Apache 2.0
Model size: 1.33 GB
Exposes embeddings? no
Tags:
detection, coco, tf2, centernet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-hg104-1024-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
centernet-hg104-512-coco-tf2#
CenterNet model from Objects as Points with the Hourglass-104 backbone trained on COCO resized to 512x512.
Details
Model name:
centernet-hg104-512-coco-tf2Model source: tensorflow/models
Model author: Xingyi Zhou, et al.
Model license: Apache 2.0
Model size: 1.49 GB
Exposes embeddings? no
Tags:
detection, coco, tf2, centernet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-hg104-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
centernet-mobilenet-v2-fpn-512-coco-tf2#
CenterNet model from Objects as Points with the MobileNetV2 backbone trained on COCO resized to 512x512.
Details
Model name:
centernet-mobilenet-v2-fpn-512-coco-tf2Model source: tensorflow/models
Model author: Xingyi Zhou, et al.
Model license: Apache 2.0
Model size: 41.98 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, centernet, mobilenet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-mobilenet-v2-fpn-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
centernet-resnet101-v1-fpn-512-coco-tf2#
CenterNet model from Objects as Points with the ResNet-101v1 backbone + FPN trained on COCO resized to 512x512.
Details
Model name:
centernet-resnet101-v1-fpn-512-coco-tf2Model source: tensorflow/models
Model author: Xingyi Zhou, et al.
Model license: Apache 2.0
Model size: 329.96 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, centernet, resnet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-resnet101-v1-fpn-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
centernet-resnet50-v1-fpn-512-coco-tf2#
CenterNet model from Objects as Points with the ResNet-50-v1 backbone + FPN trained on COCO resized to 512x512.
Details
Model name:
centernet-resnet50-v1-fpn-512-coco-tf2Model source: tensorflow/models
Model author: Xingyi Zhou, et al.
Model license: Apache 2.0
Model size: 194.61 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, centernet, resnet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-resnet50-v1-fpn-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
centernet-resnet50-v2-512-coco-tf2#
CenterNet model from Objects as Points with the ResNet-50v2 backbone trained on COCO resized to 512x512.
Details
Model name:
centernet-resnet50-v2-512-coco-tf2Model source: tensorflow/models
Model author: Xingyi Zhou, et al.
Model license: Apache 2.0
Model size: 226.95 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, centernet, resnet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-resnet50-v2-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
deeplabv3-cityscapes-tf#
DeepLabv3+ semantic segmentation model from Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation with Xception backbone trained on the Cityscapes dataset.
Details
Model name:
deeplabv3-cityscapes-tfModel source: tensorflow/models
Model author: Liang-Chieh Chen, et al.
Model license: Apache 2.0
Model size: 158.04 MB
Exposes embeddings? no
Tags:
segmentation, cityscapes, tf, deeplabv3
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("deeplabv3-cityscapes-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
deeplabv3-mnv2-cityscapes-tf#
DeepLabv3+ semantic segmentation model from Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation with MobileNetV2 backbone trained on the Cityscapes dataset.
Details
Model name:
deeplabv3-mnv2-cityscapes-tfModel source: tensorflow/models
Model author: Liang-Chieh Chen, et al.
Model license: Apache 2.0
Model size: 8.37 MB
Exposes embeddings? no
Tags:
segmentation, cityscapes, tf, deeplabv3
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("deeplabv3-mnv2-cityscapes-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d0-512-coco-tf2#
EfficientDet-D0 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 512x512.
Details
Model name:
efficientdet-d0-512-coco-tf2Model source: tensorflow/models
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 29.31 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d0-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d0-coco-tf1#
EfficientDet-D0 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.
Details
Model name:
efficientdet-d0-coco-tf1Model source: voxel51/automl
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 38.20 MB
Exposes embeddings? no
Tags:
detection, coco, tf1, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=1.14,<2
GPU support
yes
Packages:
tensorflow-gpu>=1.14,<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d0-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d1-640-coco-tf2#
EfficientDet-D1 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 640x640.
Details
Model name:
efficientdet-d1-640-coco-tf2Model source: tensorflow/models
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 49.44 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d1-640-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d1-coco-tf1#
EfficientDet-D1 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.
Details
Model name:
efficientdet-d1-coco-tf1Model source: voxel51/automl
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 61.64 MB
Exposes embeddings? no
Tags:
detection, coco, tf1, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=1.14,<2
GPU support
yes
Packages:
tensorflow-gpu>=1.14,<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d1-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d2-768-coco-tf2#
EfficientDet-D2 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 768x768.
Details
Model name:
efficientdet-d2-768-coco-tf2Model source: tensorflow/models
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 60.01 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d2-768-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d2-coco-tf1#
EfficientDet-D2 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.
Details
Model name:
efficientdet-d2-coco-tf1Model source: voxel51/automl
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 74.00 MB
Exposes embeddings? no
Tags:
detection, coco, tf1, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=1.14,<2
GPU support
yes
Packages:
tensorflow-gpu>=1.14,<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d2-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d3-896-coco-tf2#
EfficientDet-D3 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 896x896.
Details
Model name:
efficientdet-d3-896-coco-tf2Model source: tensorflow/models
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 88.56 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d3-896-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d3-coco-tf1#
EfficientDet-D3 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.
Details
Model name:
efficientdet-d3-coco-tf1Model source: voxel51/automl
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 106.44 MB
Exposes embeddings? no
Tags:
detection, coco, tf1, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=1.14,<2
GPU support
yes
Packages:
tensorflow-gpu>=1.14,<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d3-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d4-1024-coco-tf2#
EfficientDet-D4 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 1024x1024.
Details
Model name:
efficientdet-d4-1024-coco-tf2Model source: tensorflow/models
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 151.15 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d4-1024-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d4-coco-tf1#
EfficientDet-D4 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.
Details
Model name:
efficientdet-d4-coco-tf1Model source: voxel51/automl
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 175.33 MB
Exposes embeddings? no
Tags:
detection, coco, tf1, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=1.14,<2
GPU support
yes
Packages:
tensorflow-gpu>=1.14,<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d4-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d5-1280-coco-tf2#
EfficientDet-D5 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 1280x1280.
Details
Model name:
efficientdet-d5-1280-coco-tf2Model source: tensorflow/models
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 244.41 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d5-1280-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d5-coco-tf1#
EfficientDet-D5 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.
Details
Model name:
efficientdet-d5-coco-tf1Model source: voxel51/automl
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 275.81 MB
Exposes embeddings? no
Tags:
detection, coco, tf1, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=1.14,<2
GPU support
yes
Packages:
tensorflow-gpu>=1.14,<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d5-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d6-1280-coco-tf2#
EfficientDet-D6 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 1280x1280.
Details
Model name:
efficientdet-d6-1280-coco-tf2Model source: tensorflow/models
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 375.63 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d6-1280-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d6-coco-tf1#
EfficientDet-D6 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.
Details
Model name:
efficientdet-d6-coco-tf1Model source: voxel51/automl
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 416.43 MB
Exposes embeddings? no
Tags:
detection, coco, tf1, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=1.14,<2
GPU support
yes
Packages:
tensorflow-gpu>=1.14,<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d6-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
efficientdet-d7-1536-coco-tf2#
EfficientDet-D7 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 1536x1536.
Details
Model name:
efficientdet-d7-1536-coco-tf2Model source: tensorflow/models
Model author: Mingxing Tan, et al.
Model license: Apache 2.0
Model size: 376.20 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, efficientdet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d7-1536-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
faster-rcnn-inception-resnet-atrous-v2-coco-tf#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks atrous version with Inception backbone trained on COCO.
Details
Model name:
faster-rcnn-inception-resnet-atrous-v2-coco-tfModel source: tensorflow/models
Model author: Shaoqing Ren, et al.
Model license: Apache 2.0
Model size: 234.46 MB
Exposes embeddings? no
Tags:
detection, coco, tf, faster-rcnn, inception, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-inception-resnet-atrous-v2-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
faster-rcnn-inception-resnet-atrous-v2-lowproposals-coco-tf#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks atrous version with low-proposals and Inception backbone trained on COCO.
Details
Model name:
faster-rcnn-inception-resnet-atrous-v2-lowproposals-coco-tfModel source: tensorflow/models
Model author: Shaoqing Ren, et al.
Model license: Apache 2.0
Model size: 234.46 MB
Exposes embeddings? no
Tags:
detection, coco, tf, faster-rcnn, inception, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-inception-resnet-atrous-v2-lowproposals-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
faster-rcnn-inception-v2-coco-tf#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with Inception v2 backbone trained on COCO.
Details
Model name:
faster-rcnn-inception-v2-coco-tfModel source: tensorflow/models
Model author: Shaoqing Ren, et al.
Model license: Apache 2.0
Model size: 52.97 MB
Exposes embeddings? no
Tags:
detection, coco, tf, faster-rcnn, inception
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-inception-v2-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
faster-rcnn-nas-coco-tf#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with NAS-net backbone trained on COCO.
Details
Model name:
faster-rcnn-nas-coco-tfModel source: tensorflow/models
Model author: Shaoqing Ren, et al.
Model license: Apache 2.0
Model size: 404.95 MB
Exposes embeddings? no
Tags:
detection, coco, tf, faster-rcnn
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-nas-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
faster-rcnn-nas-lowproposals-coco-tf#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with low-proposals and NAS-net backbone trained on COCO.
Details
Model name:
faster-rcnn-nas-lowproposals-coco-tfModel source: tensorflow/models
Model author: Shaoqing Ren, et al.
Model license: Apache 2.0
Model size: 404.88 MB
Exposes embeddings? no
Tags:
detection, coco, tf, faster-rcnn
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-nas-lowproposals-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
faster-rcnn-resnet101-coco-tf#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with ResNet-101 backbone trained on COCO.
Details
Model name:
faster-rcnn-resnet101-coco-tfModel source: tensorflow/models
Model author: Shaoqing Ren, et al.
Model license: Apache 2.0
Model size: 186.41 MB
Exposes embeddings? no
Tags:
detection, coco, tf, faster-rcnn, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet101-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
faster-rcnn-resnet101-lowproposals-coco-tf#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with low-proposals and ResNet-101 backbone trained on COCO.
Details
Model name:
faster-rcnn-resnet101-lowproposals-coco-tfModel source: tensorflow/models
Model author: Shaoqing Ren, et al.
Model license: Apache 2.0
Model size: 186.41 MB
Exposes embeddings? no
Tags:
detection, coco, tf, faster-rcnn, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet101-lowproposals-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
faster-rcnn-resnet50-coco-tf#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with ResNet-50 backbone trained on COCO.
Details
Model name:
faster-rcnn-resnet50-coco-tfModel source: tensorflow/models
Model author: Shaoqing Ren, et al.
Model license: Apache 2.0
Model size: 113.57 MB
Exposes embeddings? no
Tags:
detection, coco, tf, faster-rcnn, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet50-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
faster-rcnn-resnet50-lowproposals-coco-tf#
Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with low-proposals and ResNet-50 backbone trained on COCO.
Details
Model name:
faster-rcnn-resnet50-lowproposals-coco-tfModel source: tensorflow/models
Model author: Shaoqing Ren, et al.
Model license: Apache 2.0
Model size: 113.57 MB
Exposes embeddings? no
Tags:
detection, coco, tf, faster-rcnn, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet50-lowproposals-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
inception-resnet-v2-imagenet-tf1#
Inception v2 model from Rethinking the Inception Architecture for Computer Vision trained on ImageNet.
Details
Model name:
inception-resnet-v2-imagenet-tf1Model source: tensorflow/models
Model author: Christian Szegedy, et al.
Model license: Apache 2.0
Model size: 213.81 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, tf1, inception, resnet
Requirements
CPU support
yes
Packages:
tensorflow<2
GPU support
yes
Packages:
tensorflow-gpu<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("inception-resnet-v2-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
inception-v4-imagenet-tf1#
Inception v4 model from Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning trained on ImageNet.
Details
Model name:
inception-v4-imagenet-tf1Model source: tensorflow/models
Model author: Christian Szegedy, et al.
Model license: Apache 2.0
Model size: 163.31 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, tf1, inception
Requirements
CPU support
yes
Packages:
tensorflow<2
GPU support
yes
Packages:
tensorflow-gpu<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("inception-v4-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
mask-rcnn-inception-resnet-v2-atrous-coco-tf#
Mask R-CNN model from Mask R-CNN atrous version with Inception backbone trained on COCO.
Details
Model name:
mask-rcnn-inception-resnet-v2-atrous-coco-tfModel source: tensorflow/models
Model author: Kaiming He, et al.
Model license: Apache 2.0
Model size: 254.51 MB
Exposes embeddings? no
Tags:
instances, coco, tf, mask-rcnn, inception, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-inception-resnet-v2-atrous-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
mask-rcnn-inception-v2-coco-tf#
Mask R-CNN model from Mask R-CNN with Inception backbone trained on COCO.
Details
Model name:
mask-rcnn-inception-v2-coco-tfModel source: tensorflow/models
Model author: Kaiming He, et al.
Model license: Apache 2.0
Model size: 64.03 MB
Exposes embeddings? no
Tags:
instances, coco, tf, mask-rcnn, inception
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-inception-v2-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
mask-rcnn-resnet101-atrous-coco-tf#
Mask R-CNN model from Mask R-CNN atrous version with ResNet-101 backbone trained on COCO.
Details
Model name:
mask-rcnn-resnet101-atrous-coco-tfModel source: tensorflow/models
Model author: Kaiming He, et al.
Model license: Apache 2.0
Model size: 211.56 MB
Exposes embeddings? no
Tags:
instances, coco, tf, mask-rcnn, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-resnet101-atrous-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
mask-rcnn-resnet50-atrous-coco-tf#
Mask R-CNN model from Mask R-CNN atrous version with ResNet-50 backbone trained on COCO.
Details
Model name:
mask-rcnn-resnet50-atrous-coco-tfModel source: tensorflow/models
Model author: Kaiming He, et al.
Model license: Apache 2.0
Model size: 138.29 MB
Exposes embeddings? no
Tags:
instances, coco, tf, mask-rcnn, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-resnet50-atrous-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
mobilenet-v2-imagenet-tf1#
MobileNetV2 model from MobileNetV2: Inverted Residuals and Linear Bottlenecks trained on ImageNet.
Details
Model name:
mobilenet-v2-imagenet-tf1Model source: tensorflow/models
Model author: Mark Sandler, et al.
Model license: Apache 2.0
Model size: 13.64 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, tf1, mobilenet
Requirements
CPU support
yes
Packages:
tensorflow<2
GPU support
yes
Packages:
tensorflow-gpu<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("mobilenet-v2-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
resnet-v1-50-imagenet-tf1#
ResNet-50 v1 model from Deep Residual Learning for Image Recognition trained on ImageNet.
Details
Model name:
resnet-v1-50-imagenet-tf1Model source: tensorflow/models
Model author: Kaiming He, et al.
Model license: Apache 2.0
Model size: 97.84 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, tf1, resnet
Requirements
CPU support
yes
Packages:
tensorflow<2
GPU support
yes
Packages:
tensorflow-gpu<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("resnet-v1-50-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
resnet-v2-50-imagenet-tf1#
ResNet-50 v2 model from Deep Residual Learning for Image Recognition trained on ImageNet.
Details
Model name:
resnet-v2-50-imagenet-tf1Model source: tensorflow/models
Model author: Kaiming He, et al.
Model license: Apache 2.0
Model size: 97.86 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, tf1, resnet
Requirements
CPU support
yes
Packages:
tensorflow<2
GPU support
yes
Packages:
tensorflow-gpu<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("resnet-v2-50-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
rfcn-resnet101-coco-tf#
R-FCN object detection model from R-FCN: Object Detection via Region-based Fully Convolutional Networks with ResNet-101 backbone trained on COCO.
Details
Model name:
rfcn-resnet101-coco-tfModel source: tensorflow/models
Model author: Jifeng Dai, et al.
Model license: Apache 2.0
Model size: 208.16 MB
Exposes embeddings? no
Tags:
detection, coco, tf, rfcn, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("rfcn-resnet101-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
ssd-inception-v2-coco-tf#
Inception Single Shot Detector model from SSD: Single Shot MultiBox Detector trained on COCO.
Details
Model name:
ssd-inception-v2-coco-tfModel source: tensorflow/models
Model author: Wei Liu, et al.
Model license: Apache 2.0
Model size: 97.50 MB
Exposes embeddings? no
Tags:
detection, coco, tf, ssd, inception
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-inception-v2-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
ssd-mobilenet-v1-coco-tf#
Single Shot Detector model from SSD: Single Shot MultiBox Detector with MobileNetV1 backbone trained on COCO.
Details
Model name:
ssd-mobilenet-v1-coco-tfModel source: tensorflow/models
Model author: Wei Liu, et al.
Model license: Apache 2.0
Model size: 27.83 MB
Exposes embeddings? no
Tags:
detection, coco, tf, ssd, mobilenet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-mobilenet-v1-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
ssd-mobilenet-v1-fpn-640-coco17#
MobileNetV1 model from MobileNetV2: Inverted Residuals and Linear Bottlenecks resized to 640x640.
Details
Model name:
ssd-mobilenet-v1-fpn-640-coco17Model source: tensorflow/models
Model author: Mark Sandler, et al.
Model license: Apache 2.0
Model size: 43.91 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, ssd, mobilenet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-mobilenet-v1-fpn-640-coco17")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
ssd-mobilenet-v1-fpn-coco-tf#
FPN Single Shot Detector model from SSD: Single Shot MultiBox Detector with MobileNetV1 backbone trained on COCO.
Details
Model name:
ssd-mobilenet-v1-fpn-coco-tfModel source: tensorflow/models
Model author: Wei Liu, et al.
Model license: Apache 2.0
Model size: 48.97 MB
Exposes embeddings? no
Tags:
detection, coco, tf, ssd, mobilenet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-mobilenet-v1-fpn-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
ssd-mobilenet-v2-320-coco17#
MobileNetV2 model from MobileNetV2: Inverted Residuals and Linear Bottlenecks resized to 320x320.
Details
Model name:
ssd-mobilenet-v2-320-coco17Model source: tensorflow/models
Model author: Mark Sandler, et al.
Model license: Apache 2.0
Model size: 43.91 MB
Exposes embeddings? no
Tags:
detection, coco, tf2, ssd, mobilenet
Requirements
CPU support
yes
Packages:
tensorflow>=2|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-mobilenet-v2-320-coco17")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
ssd-resnet50-fpn-coco-tf#
FPN Single Shot Detector model from SSD: Single Shot MultiBox Detector with ResNet-50 backbone trained on COCO.
Details
Model name:
ssd-resnet50-fpn-coco-tfModel source: tensorflow/models
Model author: Wei Liu, et al.
Model license: Apache 2.0
Model size: 128.07 MB
Exposes embeddings? no
Tags:
detection, coco, tf, ssd, resnet
Requirements
CPU support
yes
Packages:
tensorflow|tensorflow-macos
GPU support
yes
Packages:
tensorflow-gpu|tensorflow>=2|tensorflow-macos
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-resnet50-fpn-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
vgg16-imagenet-tf1#
VGG-16 model from Very Deep Convolutional Networks for Large-Scale Image Recognition trained on ImageNet.
Details
Model name:
vgg16-imagenet-tf1Model source: https://gist.github.com/ksimonyan/211839e770f7b538e2d8#file-readme-md
Model author: Karen Simonyan, et al.
Model license: CC-BY-4.0
Model size: 527.80 MB
Exposes embeddings? yes
Tags:
classification, embeddings, logits, imagenet, tf1, vgg
Requirements
CPU support
yes
Packages:
tensorflow<2
GPU support
yes
Packages:
tensorflow-gpu<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "imagenet-sample",
6 dataset_name=fo.get_default_dataset_name(),
7 max_samples=50,
8 shuffle=True,
9)
10
11model = foz.load_zoo_model("vgg16-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)
yolo-v2-coco-tf1#
YOLOv2 model from YOLO9000: Better, Faster, Stronger trained on COCO.
Details
Model name:
yolo-v2-coco-tf1Model source: thtrieu/darkflow
Model author: Joseph Redmon, et al.
Model license: GPL-3.0
Model size: 194.49 MB
Exposes embeddings? no
Tags:
detection, coco, tf1, yolo
Requirements
CPU support
yes
Packages:
tensorflow<2
GPU support
yes
Packages:
tensorflow-gpu<2
Example usage
1import fiftyone as fo
2import fiftyone.zoo as foz
3
4dataset = foz.load_zoo_dataset(
5 "coco-2017",
6 split="validation",
7 dataset_name=fo.get_default_dataset_name(),
8 max_samples=50,
9 shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo-v2-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)